There are many different things that can cause us to change what we value. Some seem like change-processes that our current values would endorse:

• I thought a lot about the issue, decided that two of my values were in conflict, and chose to prioritize one over the other.

Some seem like random processes we should resist:

• I got hit in the head, and the damage caused me to change my personality and values.

Some seem actively adversarial:

• Years of propaganda wore me down and caused me to love Dear Leader.

Or self-deceptive:

• I subconsciously realized that opinion X was high status, and found it expedient to adopt this as well.

In general I think people's opinions on the issue depend on how common they think these different cases are. I am generally quite pessimistic here; I think the first case is quite rare, and most cases that appear to be of this form are really examples of the third or fourth case. This makes me pessimistic about the long-term future, and I am interested in what we can do to reduce the influence of the last three cases.

How bad (or possibly good) value drift and lifestyle drift are will depend your definition of the phenomenon, as you acknowledge yourself.

Here's how I think about the topic:

I use the terms value drift and lifestyle drift in a broad sense to mean internal or external changes that would lead you to lose most of the expected altruistic value of your life. Value drift is internal; it describes changes to your value system or motivation. Lifestyle drift is external; the term captures changes in your life circumstances leading to difficulties implementing your values. Internally, value drift could occur by ceasing to see helping others as one of your life’s priorities (losing the ‘A’ in EA), or loosing the motivation to work on the highest-priority cause areas or interventions (losing the ‘E’ in EA). Externally, lifestyle drift could occur (as described in Joey's post) by giving up a substantial fraction of your effectively altruistic resources for non-effectively altruistic purposes, thus reducing your capacity to do good. Concretely, this could involve deciding to spend a lot of money on buying a (larger) house, having a (fancier) wedding, traveling around the world (more frequently or expensively), etc. Quoting a previous article on the topic:

Of course, changing your cause area or intervention to something that is equally or more effective within the EA framework does not count as value drift. Note that even if your future self were to decide to leave the EA community, as long as you still see ‘helping others effectively’ as one of your top-priorities in life it might not constitute value drift. (...)

Most of the potential value of EAs lies in the mid- to long-term, when more and more people in the community take up highly effective career paths and build their professional expertise to reach their ‘peak productivity’ (likely in their 40s). If value drift is common, then many of the people currently active in the community will cease to be interested in doing the most good long before they reach this point. This is why, speaking for myself, losing my altruistic motivation in the future would equal a small moral tragedy to my present self. I think that as EAs we can reasonably have a preference for our future selves not to abandon our fundamental commitment to altruism or effectiveness.

Value drift towards the right values (i.e. Effective Altruism) is good, value drift away from them is bad. Value drift among EAs is likely to be a bad thing due to regression to the mean. We can imagine better values within EA, but there's no reason to expect value drift to go in the right direction. Of course we can identify better values and promote them among EAs, but that seems notably distinct from value drift.

On the other hand, we can imagine people with bad values who should regress to the mean, and would encourage value drift there.

My sense is that there are definitely divergent and convergent value drifts. That is to say, some people seem to become less coherent over time and others more coherent.

It's a convergent instrumental goal to preserve one's values. If you change your goals / values, you will generally achieve / fulfill them less.

Value-drifting someone else might be positive for you, at least if you only consider the first-order consequences, but it generally seems pretty unvirtuous and uncooperative to me. A world where value-drifting people is socially acceptable is probably worse than a world where it's not.